Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Interaction of G protein-coupled receptors (GPCR) with ligands results in conformational changes in the receptor structure that enables its interaction with specific classes of heterotrimeric G proteins (Gudermann et al., 1997; Bockaert and Pin, 1999). The activated G protein subunits  and are then able to modulate the activity of downstream effectors. A single GPCR can interact with several distinct G protein combinations that exist in a given cell, thereby generating divergent signaling through a single receptor subtype (Kenakin, 1995). How this selectivity is achieved in terms of protein-protein interactions is not well understood; this is mainly due to the lack of structural data on receptor-G protein interaction domains. Numerous mutagenesis and biochemical studies (Liu et al., 1995; Kostenis et al., 1997) have shown that the carboxy-terminal portion of G subunits is an important determinant of GPCR contact specificity. Synthetic C-terminal peptides of transducin (G_t) and monoclonal antibodies specific for the G_t carboxy-terminal portion prevent the interaction between transducin and rhodopsin (Hamm et al., 1988; Mazzoni et al., 1991). Construction of a chimeric G_q protein by substitution of its three C-terminal amino acids by those of a G_i2 protein switches GPCR specificity from the adenylyl cyclase to the phospholipase C pathway (Conklin et al., 1993). The use of such chimeric G proteins, exchanging up to nine amino acids of their extreme C-terminal portion, has been reported to direct GPCRs differing in their G protein-coupling specificity toward common effector systems, such as the production of inositol phosphates or the mobilization of intracellular Ca²⁺ (Milligan and Rees, 1999).

The _2A-adrenoceptor (_2A-AR; receptor classification, 2.1.ADR.A2A) has been shown to activate multiple and distinct effector pathways, such as inhibition and activation of adenylyl cyclase (Fraser et al., 1989; Eason et al., 1992), activation of phospholipase C (Cotecchia et al., 1990; Dorn et al., 1997), activation of K⁺ channels (Fraser et al., 1989), and inhibition of Ca²⁺ channels (Airriess et al., 1997). The dual signaling properties of _2A-AR to the inhibition and activation of adenylyl cyclase is dependent on the ligand structure and is mediated by two distinct G proteins of the G_i/o and G_s families (Eason et al., 1994; Eason and Liggett, 1996; Brink et al., 2000). The aim of the current study was to examine the exact contribution of each of the last five carboxy-terminal amino acids of the G_s protein in divergent _2A-AR signaling and their effect on _2A-AR ligand binding properties. Therefore, a collection of mutant G_o proteins in which the last five amino acids positions were systematically exchanged between G_o and G_s proteins was constructed. Functional analysis was performed by co-expression of the mutant G_o proteins with a wt _2A-AR. Agonist-dependent binding of the stable GTP analog [³⁵S]GTPS to the mutant G_o proteins, and the binding of both ³H-agonist and ³H-antagonist to the _2A-AR co-expressed in COS-7 cellular membranes were measured. Because none of the mutant G_o proteins contained an ADP-ribosylation site by Bordetella pertussis toxin (PTX), cells were treated with PTX to avoid _2A-AR coupling to endogenous G_i/o proteins in COS-7 cells. The G_o protein-derived Gly residue at the 3 position away from the protein C-terminal extremity demonstrated a pivotal role in decreasing the efficacy of the partial agonist clonidine. Moreover, a similar mutation induced a decrease in agonist, but not antagonist, binding affinity to the _2A-AR. These results are discussed in view of structural conformation data on G protein C-terminal portion and receptor interactions.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Construction of Mutant G_o Proteins. The investigated mutant G_o proteins were generated by PCR on linearized pCR3.1/G_o cDNA plasmid (Pauwels et al., 2001) using a sense primer designed according to the rat G_o cDNA nucleotide sequence (GenBank accession number M17526) and a mutagenic reverse primer carrying the respective mutation; their sequences are indicated in Table 1. The amplification conditions were similar, as previously described (Pauwels et al., 2001). PCR products were cloned into a pCR3.1 expression vector (Invitrogen, San Diego, CA) and fully sequenced on an ABI 310 Prism genetic analyzer (PerkinElmer Life Science Products, Foster City, CA) using a Big Dye Terminator cycle sequencing ready reaction kit (PerkinElmer Life Science Products), confirming the presence of the respective mutations.

View this table: [in this window] [in a new window]   TABLE 1 Sequence characteristics of the C-terminal portion of the mutant Go proteins The last six C-terminal amino acids of the rat Go protein (Arg349 to Tyr354) were exchanged with the equivalent residues of either the rat Gs or the mouse G15 protein. Systematic mutations of the four residues, diverging between the C-terminal portions of Go and Gs proteins, were realized as described under Experimental Procedures using the indicated mutagenic reverse primers. The arrow indicates the position of the PTX-mediated ADP ribosylation site in the wt Go protein. The nucleotides or amino acids indicated in bold are those that are modified according to the Go/s cDNA sequence. The -6 residue is identical for the Go and Gs proteins (Arg) but different for the G15 protein (Asp). It will only be indicated for mutants involving this latter G protein.

Cell Culture and Transfection Procedures. The COS-7 cell line (ATCC: CRL 1651; American Type Culture Collection, Manassas, VA) was cultured in Petri dishes (50 cm²) with Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal calf serum. Cells grown to 60 to 80% confluence were used for transfection using a LipofectAMINE plus kit (Invitrogen, Paisley, UK). pCR3.1 plasmid (3-0.03 µg) containing the wt human _2A-AR gene (receptor classification, 2.1.ADR.A2A; GenBank accession number M23533) and 3 µg of either empty plasmid or indicated mutant G_o protein plasmid were mixed with 10 µl of LipofectAMINE plus reagent diluted in 0.2 ml of Opti-MEM and incubated at room temperature for 15 min. Subsequently, 20 µl of LipofectAMINE reagent diluted 20 times in 0.2 ml of Opti-MEM was added and incubated for 15 min. COS-7 cells were exposed to the plasmid/LipofectAMINE mixture with 5 ml of Opti-MEM for 3 h at 37°C. Thereafter, cells were incubated with 10 ml of complete growth medium and harvested 48 h after transfection. Treatment with PTX (20 ng/ml) was performed overnight before membranes were prepared.

Membrane Preparation and Radioligand Binding Experiments. Membrane preparation steps were performed at 4°C. Cells were washed with phosphate-buffered saline and stored at 80°C. Cells were scraped mechanically in 10 mM Tris-HCl, 0.1 M EDTA, pH 7.5, and centrifuged for 10 min at 45,000g. The pellet was homogenized in the same buffer and centrifuged under similar conditions. The final pellet was distributed at 0.5 to 1.5 mg of protein/ml in Tris-EDTA buffer and stored at 80°C. Membrane preparations were diluted in 20 mM Hepes, 100 mM NaCl, 3 mM MgCl₂, and 0.2 mM ascorbic acid, pH 7.4, and used for the binding study with [³H]2-(2-ethoxy-2,3-dihydro-benzo[1,4]dioxin-2-yl)-4,5-dihydro-1H-imidazole ([³H]RX 821002), [³H]5-bromo-6-(2-imidazoline-2-ylamino)quinoxaline tartrate ([³H]UK 14304), and [³H]clonidine, as described (Wurch et al., 1999). Nonspecific radioligand binding was determined in the presence of 10 µM phentolamine. Scatchard analysis was performed as described (Pauwels et al., 1996) using concentrations of radioligand ranging from 0.3 to 10 nM for [³H]RX 821002, 0.2 to 100 nM for [³H]clonidine, and 0.04 to 40 nM for [³H]UK 14304. Data were analyzed by the nonlinear square curve-fitting program, Ligand version 4.0 (Biosoft, Cambridge, UK; Rovati et al., 1989)

[³⁵S]GTPS Binding Responses. Agonist-independent (basal) and agonist-dependent [³⁵S]GTPS binding responses were performed to the membrane preparations described above in 20 mM Hepes, 30 µM GDP, 100 mM NaCl, 3 mM MgCl₂, and 0.2 mM ascorbic acid, pH 7.4. Maximal stimulation of [³⁵S]GTPS binding was defined in the presence of 10 µM ()-epinephrine and calculated versus basal [³⁵S]GTPS binding, unless otherwise indicated. The maximal capacity of recombinant mutant G_o protein agonist-mediated activation was determined by saturation [³⁵S]GTPS binding on the same membrane preparations in GDP 30 µM, 0.5 nM [³⁵S]GTPS, and 0 to 300 nM unlabeled GTPS in 20 mM Hepes, 100 mM NaCl, 3 mM MgCl₂, and 0.2 mM ascorbic acid, pH 7.4. The binding reaction was terminated by rapid filtration through Whatman GF/B glass fiber filters (Brandel, Gaithersburg, MD) treated as described (Pauwels et al., 2001). EC₅₀ values were derived graphically as the concentration of compound yielding 50% of its own maximal [³⁵S]GTPS binding response. The potency of clonidine to antagonize ()-epinephrine-mediated [³⁵S]GTPS binding responses was calculated according to the equation K_B = (B)/[(A')/(A)  1], where B is the clonidine concentration, and A and A' are the EC₅₀ values of ()-epinephrine in the absence and presence of clonidine, respectively.

Immunological Detection of Mutant G_o Protein Expression. Membrane fractions of COS-7 cells transiently co-expressing the _2A-AR and mutant G_o proteins were prepared as described above. Total proteins were separated by denaturing 12.5% (w/v) SDS-polyacrylamide gel electrophoresis, as described (Laemmli, 1970). After electrophoresis, proteins were blotted onto a nylon membrane by semidry electrotransfer (23 V, 45 min) in 25 mM Tris-HCl, pH 8.3, 190 mM glycine, 20% (v/v) methanol. Proteins were probed using a monoclonal antibody raised against a peptide corresponding to amino acids 18 to 33 of the G_o protein. The incubation was performed in phosphate-buffered saline buffer containing 0.1% (w/v) Tween 20, 5% (w/v) dry nonfat milk, and the antibody at a dilution of 1:1000. Proteins were visualized with an anti-mouse IgG antibody coupled to horseradish peroxidase using a chemiluminescence reaction. Quantification of the immunodetected signal was performed using a computer-based image analysis system (Imagena 2000 software; Biocom, Les Ulis, France).

Protein Content. The protein level of membrane preparations was estimated with a dye-binding assay using a Bio-Rad kit (Bio-Rad, Hercules, CA); bovine serum albumin was used as a standard (Bradford, 1976).

Statistical Analysis. Statistical analyses were performed on K_D and B_max values of the radioligands by a one-way analysis of variance, followed by an all pairwise multiple comparison procedure (method of Tukey) between G_o/GYGLY (= G_oCys³⁵¹Tyr) and the other mutant G_o proteins.

Materials. The ABI Prism 310 genetic analyzer and big dye terminator cycle sequencing ready reaction kit were obtained from PerkinElmer Life Science Products. The Imagena 2000 software was obtained from Biocom. The pCR3.1 expression was purchased from Invitrogen. COS-7 cells were obtained from the American Type Culture Collection. The LipofectAMINE plus kit, cell culture medium, fetal calf serum, and B. pertussis toxin (50 µg/ml) were purchased from Invitrogen. [³H]RX 821002 (67 Ci/mmol), [³H]UK 14304 (74 Ci/mmol), and [³H]clonidine (70.2 Ci/mmol) were obtained from PerkinElmer Life Science Products. [³⁵S]GTPS (1035-1163 Ci/mmol) and the ECL chemiluminescence reaction kit were obtained from Amersham Pharmacia Biotech (Les Ulis, France). ()-Epinephrine and clonidine were from Sigma (St. Louis, MO).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

[³⁵S]GTPS Binding Responses as Mediated by Mutant G_o Proteins in the Co-Presence of _2A-AR. Neither wt _2A-AR nor G_oCys³⁵¹Tyr protein, expressed independently in COS-7 cells, displayed a detectable [³⁵S]GTPS binding response upon stimulation by 10 µM ()-epinephrine (not shown). Co-expression of the _2A-AR with a chimeric G_o protein in which the last five amino acids of the wt G_o protein were replaced by the equivalent portion of the G_s protein (G_o/QYELL) resulted in a low-magnitude clonidine-mediated [³⁵S]GTPS binding response (14% stimulation versus ()-epinephrine; Table 2) compared with the PTX-resistant G_oCys³⁵¹Ile protein (73% stimulation versus ()-epinephrine; Table 2). The ()-epinephrine-mediated [³⁵S]GTPS binding response was only decreased 2-fold compared with its basal [³⁵S]GTPS binding level for these two mutant G_o proteins (Table 2). Both G_oCys³⁵¹Ile and G_o/QYELL proteins differ at four amino acid positions (Gln³⁵⁰Gly, Tyr³⁵¹Ile, Glu³⁵²Gly, and Leu³⁵⁴Tyr). A gain-of-function approach to investigate which of these four amino acids may be involved in the low-magnitude profile of clonidine at the G_o/QYELL protein was conducted by measuring [³⁵S]GTPS binding responses; the data are summarized in Table 2. Basal [³⁵S]GTPS binding responses for most of the mutant G_o proteins were between 105 and 190 fmol/mg of protein; a trend for an elevated basal level was observed for the mutant G_o/QIGLL and G_o/QIGLY proteins. ()-Epinephrine (10 µM) stimulated the binding of [³⁵S]GTPS from 175 to 534%. Clonidine yielded two types of responses; it acted as a partial to efficacious agonist with maximal responses between 60 and 110% of that mediated by ()-epinephrine or as a weak agonist with a maximal response below 15% compared with ()-epinephrine. Both responses could be associated with a single amino acid position in the C-terminal portion of the G_o protein. Clonidine behaved as an efficacious agonist, with a maximal response as high as that of ()-epinephrine when a Gly residue is present at the 3 position of the mutant G_o protein (Table 2). In contrast, when the 3 position was a Glu, the maximal stimulation of [³⁵S]GTPS binding by clonidine was below 15% compared with ()-epinephrine (Table 2). The amino acid at the 5 position also influenced, but to a lesser extent, the level of mutant G_o protein activation by the clonidine-occupied _2A-AR; the 5 Gln/3 Gly combination generated the highest activation level (G_o/QIGLL, G_o/QIGLY, G_o/QYGLL, and G_o/QYGLY proteins, 83 to 110%; Table 2) compared with ()-epinephrine, whereas the mutants carrying the 5 Gly/3 Glu combination yielded almost no stimulation with clonidine (G_o/GYELY, G_o/GYELL, G_o/GIELL, and G_o/GIELY proteins, 5 to 7%; Table 2). The 1 position (Leu or Tyr) did not influence the G protein activation level independently of the other amino acid positions. Thus, four different classes based on the (5)/(3) amino acid positions in the mutant G_o proteins could be differentiated according to the rank order of their clonidine-mediated [³⁵S]GTPS binding response: G_o/Q(I/Y)GL(L/Y) > G_o/G(I/Y)GL(L/Y) G_o/Q(I/Y)EL(L/Y) > G_o/G(I/Y)EL(L/Y), as depicted in Table 2 (in bold the 5 Gly/Gln and the 3 Gly/Glu positions).

View this table: [in this window] [in a new window]   TABLE 2 [35S]GTPS binding responses of 2A-AR co-expressed with a series of mutant Go proteins Classification was performed according to the maximal [35S]GTPS binding response of clonidine calculated in percentage vs. ()-epinephrine. Co-expression of 2A-AR and respective mutant Go protein was performed as described under Experimental Procedures. All conditions were treated with PTX (20 ng/ml). Basal, 10 µM ()-epinephrine, and 10 µM clonidine-stimulated [35S]GTPS binding responses, mediated by the 2A-AR, were performed as described under Experimental Procedures. Data represent mean values ± S.E.M. of three to seven independent transfection experiments, each performed in duplicate. The bold amino acids correspond to those that are different between Go/s (Go/QYELL) and the various mutant Go proteins.

Analysis of Agonist-Occupied _2A-AR-Mediated Maximal [³⁵S]GTPS Binding Capacity to Mutant G_o Proteins. To further analyze the influence of the 5/3 amino acid composition of the G_o protein on ligand-dependent _2A-AR activation, mutant G_o/GYELL, G_o/QYGLL, G_o/GYGLL, G_o/QYELL (= G_o/s), and G_o/GYGLY (= G_oCys³⁵¹Tyr) proteins were selected to perform agonist-specific [³⁵S]GTPS binding analyses. To exclude putative differences in functional responses due to variation in the expression of the mutant G_o proteins, immunological detection indicated the expression level of the mutant G_o proteins varied between 53 and 163% compared with that of the mutant G_oCys³⁵¹Tyr protein (Fig. 1). No relation between the mutant G_o protein expression level and the clonidine-mediated maximal [³⁵S]GTPS binding response was apparent. ()-Epinephrine (10 µM)-mediated saturation [³⁵S]GTPS binding indicated a single population of high-affinity [³⁵S]GTPS binding sites for each of the investigated mutant G_o proteins. The apparent dissociation constant of [³⁵S]GTPS was not statistically different, with the exception of the G_o/GYELL protein, which yielded about a 5-fold increased K_D value (Table 3). The maximal ()-epinephrine-mediated [³⁵S]GTPS binding capacity varied between 3.97 and 11.37 pmol/mg of protein for these mutant G_o proteins. Clonidine (10 µM) stimulated [³⁵S]GTPS binding to the mutant G_o/GYGLY, G_o/QYGLL, and G_o/GYGLL proteins to the same extent as ()-epinephrine, but the mutant G_o/QYELL and G_o/GYELL proteins were only weakly stimulated, and consequently saturation analysis was not performed. Dose-dependent [³⁵S]GTPS binding response curves for ()-epinephrine yielded a 24- and 54-fold decreased potency at the _2A-AR in the co-presence of the mutant G_o/QYELL and G_o/GYELL proteins, respectively, compared with the G_oCys³⁵¹Tyr protein (Fig. 2). Clonidine potently (EC₅₀, 13.0 to 32.0 nM) stimulated [³⁵S]GTPS binding responses at the mutant G_o proteins carrying a 3 glycine residue, but it acted as a competitive antagonist of the ()-epinephrine-mediated [³⁵S]GTPS binding response at the _2A-AR in the co-presence of those mutant G_o proteins with a 3 Glu residue (Fig. 2). To evaluate the influence of putative spare _2A-ARs on mutant G_o/QYGLL and G_o/GYELL protein activation, [³⁵S]GTPS binding responses were monitored in the presence of decreasing amounts of _2A-ARs (21.1 to 0.29 pmol/mg of protein; Table 4). Decreasing the _2A-AR expression by about 50-times yielded only a slight decrease (2- to 3-fold) in potency for both ()-epinephrine and clonidine at the mutant G_o/QYGLL protein, without altering the maximal response of clonidine (Table 4). The degree of basal [³⁵S]GTPS binding was not affected by the expression level of _2A-AR (Table 4).

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Immunological detection of mutant Go protein expression in COS-7 cells in the co-presence of 2A-AR. One hundred micrograms of total cellular membrane proteins of COS-7 cells co-expressing 2A-AR and empty plasmid (A), mutant GoCys351Tyr (B), Go/QYELL (C), Go/GYELL (D), Go/QYGLL (E), and Go/GYGLL (F) proteins were separated by 12.5% SDS- polyacrylamide gel electrophoresis, blotted onto a nylon membrane, and the immunodetection was performed, as described under Experimental Procedures, using a selective anti-Go antibody. The arrow indicates a signal corresponding to the mutant Go proteins. Quantification (percentage versus mutant GoCys351Tyr protein) of the immunodetected signal was 78, 53, 92, and 167 for lane C to F, respectively.

View this table: [in this window] [in a new window]   TABLE 3 Dissociation constants and Bmax values for binding of [35S]GTPS to membrane preparations of COS-7 cells expressing 2A-AR and various mutant Go proteins Coexpression of 2A-AR and respective mutant Go protein was performed, as described under Experimental Procedures. All conditions were treated with PTX (20 ng/ml). Saturation [35S]GTPS binding responses mediated by the 2A-AR were performed as described under Experimental Procedures. Membranes were incubated with 0.5 nM [35S]GTPS, 30 µM GDP, and either without or with 0.1 to 300 nM unlabeled GTPS. KD (nM) and Bmax (pmol/mg of protein) values were deduced from saturation analysis for specific ()-epinephrine (10 µM) and/or clonidine (10 µM)-stimulated [35S]GTPS binding. Data represent mean values ± S.E.M. of four independent transfection experiments, each performed in duplicate. The bold amino acids correspond to those that are different between Go/s (Go/QYELL) and the various mutant Go proteins. Statistical analysis was performed on KD and Bmax values between Go/GYGLY and the other mutant Go proteins.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   ()-Epinephrine dose-dependent [35S]GTPS binding response curves at 2A-AR in the co-presence of various mutant Go proteins in COS-7 cells. Cultures were treated overnight with PTX (20 ng/ml) and assayed for [35S]GTPS binding, as described under Experimental Procedures. ()-Epinephrine dose-dependent response curves (A) are shown for the various mutant Go proteins (EC50, nM): Go/GYGLY (37 ± 9.2; ), Go/QYELL (900 o ± 151; ), Go/GYELL (2000 ± 94; ), Go/QYGLL (13.5 ± 2.5; ), Go/GYGLL (40.0 ± 3.5; ). Antagonism of ()-epinephrine response curves by clonidine is presented for the mutant Go/QYELL (B) and Go/GYELL proteins (C) in either the absence (closed symbols) or presence (open symbols) of clonidine (10 µM). Data are presented in percentage versus the maximal ()-epinephrine-mediated [35S]GTPS binding response for each mutant Go protein. Concentration binding curves are constructed using mean values ± S.E.M. from four independent transfection experiments, each performed in duplicate.

View this table: [in this window] [in a new window]   TABLE 4 Influence of the expression level of 2A-AR on the [35S]GTPS binding response of mutant Go/QYGLL and Go/GYELL proteins Co-transfection with 3 and 0.03 µg of 2A-AR plasmid and 3 µg of indicated mutant Go protein plasmid was performed as described under Experimental Procedures. All conditions were treated with PTX (20 ng/ml). [3H]RX 821002 (saturating concentration, 4.0 nM) and [35S]GTPS (0.5 nM) binding responses were performed as described under Experimental Procedures. Data represent mean values + S.E.M. of three independent transfection experiments, each performed in duplicate.

Saturation Radioligand Binding Responses at _2A-AR in the Co-Presence of Mutant G_o Proteins. Saturation binding experiments using an ₂ AR antagonist [³H]RX 821002, an efficacious ₂ AR agonist [³H]UK 14304 (Jasper et al., 1998), and an ₂ AR partial agonist [³H]clonidine (Jasper et al., 1998) were performed (Table 5) to assess their binding properties to the _2A-AR in either the absence or in the co-presence of the mutant G_o/GYGLY, G_o/QYELL, G_o/GYELL, G_o/QYGLL, and G_o/GYGLL proteins. The equilibrium dissociation constant and maximal binding capacity of [³H]RX 821002 at the _2A-AR were not statistically different with each of the co-expressed mutant G_o proteins, although a slightly higher amount of _2A-AR binding sites in the co-presence of the mutant G_o/QYELL protein was observed (Table 5). In contrast, the K_D values for the labeled agonists were highly dependent on the co-expressed mutant G_o protein; a 12- to 39-fold and a 9- to 33-fold increased (P < 0.05) dissociation constant value for [³H]clonidine and [³H]UK 14304, respectively, was observed for the _2A-AR in the presence of either a G_o/QYELL or G_o/GYELL protein compared with the mutant G_o proteins carrying a glycine as the 3 amino acid residue. The maximal radioligand binding capacity at the _2A-AR sites was either slightly increased (P < 0.05, [³H]clonidine) or unaffected (P > 0.05, [³H]UK 14304) by the presence of the mutant G_o proteins but was lower for both radiolabeled agonists compared with [³H]RX 821002. The absence of a difference in maximal [³H]UK 14304 binding sites for the G_o/QYELL and G_o/GYELL proteins compared with the other mutant G_o proteins containing a 3 Gly residue suggests that both of them can also exist in an _2A-AR-coupled state (Table 5). In the absence of recombinant G proteins, the binding parameters of both [³H]clonidine and [³H]UK 14304 were close to those observed for _2A-ARs in the co-presence of a mutant G_o/GYELL protein (Table 5).

View this table: [in this window] [in a new window]   TABLE 5 KD and Bmax values for the binding of [3H]RX 821002, [3H]clonidine, and [3H]UK 14304 to membrane preparations of COS-7 cells expressing the 2A-AR in either the absence or presence of various mutant Go proteins Co-expression of 2A-AR and either empty plasmid or respective mutant Go protein was performed as described under Experimental Procedures. All conditions were treated with PTX (20 ng/ml). The equilibrium dissociation constant (KD, nM) and maximal radioligand binding capacity (Bmax, pmol/mg of protein) were determined for each condition, as described under Experimental Procedures, according to a monophasic Scatchard analysis. Data represent mean values ± S.E.M. of four independent transfection experiments, each performed in duplicate. The bold amino acids correspond to those that are different between Go/s (Go/QYELL) and the various mutant Go proteins. Statistical analysis was performed on ligand's KD and Bmax values between Go/GYGLY and the other mutant Go proteins or empty plasmid.

o

15

o

o

o/15

o

o

o

o

o

o

View this table: [in this window] [in a new window]   TABLE 6 KD and Bmax values for the binding of various radioligands to membrane preparations of COS-7 cells expressing the 2A-AR and mutant Go/DEINLL and Go/DEIGLL proteins Co-expression of 2A-AR and mutant Go protein was performed as described under Experimental Procedures. All conditions were treated with PTX (20 ng/ml). Saturation [35S]GTPS binding responses mediated by the 2A-AR were performed as described. Membranes were incubated with 0.5 nM [35S]GTPS, 30 µM GDP, and either without or with 0.1 to 300 nM unlabeled GTPS. KD (nM) and Bmax (pmol/mg of protein) values were deduced from saturation analysis for specific ()-epinephrine (10 µM) and/or clonidine (10 µM)-stimulated [35S]GTPS binding. The equilibrium dissociation constant (KD, nM) and maximal radioligand binding capacity (Bmax, pmol/mg of protein) were determined for each condition as described under Experimental Procedures according to a monophasic Scatchard analysis. Data represent mean values ± S.E.M. of four independent transfection experiments, each performed in duplicate. Statistical analysis was performed on ligand's KD and Bmax values between Go/DEINLL and Go/DEIGLL proteins.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

This study demonstrates reciprocal interactions between a wt _2A-AR and a G_o protein mutated in its five carboxy-terminal amino acid residues. Analysis was conducted using saturation binding experiments of either a labeled, nonhydrolyzable analog of guanine nucleotides, [³⁵S]GTPS, as well as labeled radioligands being either efficacious or partial agonists or an antagonist. Maximal agonist-mediated saturation [³⁵S]GTPS binding responses for the various mutant G_o proteins and its comparison with the maximal antagonist [³H]RX 821002 binding capacity gives an appropriate approximation of the ratio between total _2A-AR amount and activated G_o protein capacity. Among the various mutant G_o proteins, the most significant effect on the modulation of the magnitude of maximal agonist-mediated [³⁵S]GTPS binding response was the exchange, at the carboxy-terminal end of the protein, of a 3 glutamate or asparagine residue, as derived from a G_s or G₁₅ protein, respectively, for a G_i/o protein-derived glycine. When a glycine, this position, independent of the surrounding peptidic sequence corresponding either to that of a G_s or a G₁₅ protein, yielded an enhanced maximal response for the partial agonist clonidine. A single mutation at this critical position not only modulated the ligand-occupied _2A-AR-mediated [³⁵S]GTPS binding response but also reciprocally altered the agonist binding pocket at the _2A-AR because agonist equilibrium dissociation constants were decreased. This also indicates that interaction of the 3 Gly containing mutant G_o proteins stabilized an activated _2A-AR conformational state, as suggested by the increased potency of ()-epinephrine and the enhanced dissociation constants of the labeled agonists. _2A-ARs may possess an enhanced affinity for the 3 Gly containing mutant G_o proteins, as predicted by the extended ternary complex model (Lefkowitz et al., 1993). Remarkably, mutation of the 3 residue into a negatively charged Glu residue produced an effect that is opposite to that obtained at the GPCR third intracellular loop distal portion, where the mutation of a noncharged residue by either an acidic (i.e., mutant Ala²⁹³Glu _1B AR) or basic (i.e., mutant Thr³⁷³Lys _2A-AR) amino acid generated constitutive activation by constraining a G protein-coupled state of the receptor (Pauwels and Wurch, 1998). Thus, although the GPCR third intracellular loop distal portion has been postulated to interact with a G protein C-terminal end (Kostenis et al., 1997), the exact contribution of the 3 Gly versus Glu residue cannot be foreseen.

The systematic mutation of each of the last five C-terminal amino acids of the G_o protein, either alone or in combination, emphasized a pivotal role of the 3 residue. It can be either a Gly for the G_o protein studied here or for the closely related G_i1/2/3 and G_z proteins, a charged Glu in the case of G_s, or a polar Asn for the G_q/11/15/16 proteins. The nature of this peculiar residue is such that it modulates on its own the activation level of the G protein, as mediated by efficacious and partial ₂ AR agonists, without modifying its basal activation level. Similarly, a chimeric G_o protein exchanging its last six amino acids for those of a G_z protein yielded an enhanced maximal [³⁵S]GTPS binding response for the partial agonist d-medetomidine, whereas almost no stimulation of [³⁵S]GTPS binding was obtained with a chimeric G_o/q protein (Pauwels et al., 2001). Clonidine (10 µM)-occupied _2A-ARs activated a number of high-affinity [³⁵S]GTPS binding sites similar to that of the native ₂ AR agonist ()-epinephrine in the co-presence of mutant G_o proteins containing a 3 Gly residue (i.e., G_oCys³⁵¹Tyr, G_o/QYGLL, and G_o/GYGLL proteins). Therefore, clonidine and ()-epinephrine can be considered as agonists with a similar maximal response under these experimental conditions. On the other hand, when the _2A-AR was expressed with mutant G_o proteins containing a 3 Glu residue (i.e., G_o/QYELL and G_o/GYELL proteins), clonidine at saturating concentrations (10 µM) acted not only as a very weak agonist, but it also competitively antagonized the ()-epinephrine-mediated [³⁵S]GTPS binding response. Clonidine has been reported to display a comparable antagonist potency of the ()-epinephrine-mediated [³⁵S]GTPS binding response at _2A-ARs stably expressed in HEK 293 cells (Jasper et al., 1998). This shows that, depending on the co-presence of a particular mutant G_o protein, the clonidine-occupied _2A-AR is able or not to activate the G_o protein. In the absence of efficacious G_o protein activation, clonidine can antagonize the functional response of ()-epinephrine. Our data extend the implication of this residue, which has previously been involved in the selectivity of G_q protein coupling to _2A-ARs; a single mutation (Asn³⁵⁷Gly) at the 3 position of the C-terminal portion of a G_q protein renders it responsive to an agonist-activated _2A-AR (Conklin et al., 1996). Similarly, the mutant G_q Asn³⁵⁷Gly protein efficiently coupled the G_i/o-coupled muscarinic m₂ receptor to the inositol phosphate pathway, without modification of the potency of the agonist carbachol (Liu et al., 1995; Kostenis et al., 1997).

The importance of the 3 residue has also been reported on a structural basis; NMR studies on an 11-amino-acid-long peptide corresponding to the C-terminal portion of the rod cell G_t protein (the  subunit of transducin) suggests that its disordered conformation is shifted upon light activation of rhodopsin to a highly structured helical turn, followed by an open reverse turn centered at the 3 glycine residue (Kisselev et al., 1998). Fluorescence studies also revealed that the G_t protein activation leads to a conformational change at its C-terminal portion, which may provide a structural basis for communication between a G_t protein and light-activated rhodopsin (Yang et al., 1999). The formation of a highly structured motif at the C-terminal portion of the mutant G_o proteins may favor specific interactions with the _2A-AR in which conformation has been modified upon activation by an agonist. The presence of the 3 Gly residue is likely to be necessary for an optimal protein structure because 98% of the mutant G_o proteins corresponds to the native G_o protein. The flexibility of the C-terminal portion might be affected by the Glu³⁵²Gly mutation because of the loss of a negative charge, which may be stabilized by intramolecular interactions otherwise existing in the wt G_s protein. A similar effect of the 3 Asn to Gly mutation in the chimeric G_o/15 protein and the loss of the noncharged polar moiety may suggest an unique role of the glycine residue by the absence of a side chain.

A second major observation in our study consists in the decrease of the equilibrium dissociation constant of the agonists [³H]clonidine and [³H]UK 14304, but not that of the antagonist [³H]RX 821002, for binding to the _2A-AR in the co-presence of 3 Glu-containing mutant G_o proteins. These data may be interpreted in view of a conformational change of the _2A-AR state, dependent on the co-expressed mutant G_o protein. Although the antagonist recognizes both the G protein-coupled and -uncoupled states of the _2A-AR (Kenakin, 1995), the dissociation constants of the agonists are modulated by the coupling efficiency of the mutant G_o protein to the _2A-AR. Mutant G_o proteins containing a Glu residue as the 3 amino acid (i.e., G_o/QYELL, G_o/GYELL, and G_o/DEIGLL proteins) yielded a 10- to 50- fold decreased dissociation constant for the radiolabeled agonists at the _2A-AR. These data suggest that the interaction between these mutant G_o proteins, and the _2A-AR induces tiny modifications in the binding site for the agonists UK 14304 and clonidine, whereas the interaction with the antagonist RX 821002 is unaffected. Both ₂ AR agonists contain a common imidazoline ring, which constitutes a binding domain to the _2A-AR (Salminen et al., 1999) and may therefore explain why these two ligands, apart being agonists compared with the antagonist RX 821002, are similarly affected by the mutations in the G_o protein. Thus, the data described here indicate that a single mutation in the G_o/s protein carboxy-terminal portion increased the affinity of the chimeric G protein for the _2A-AR. Several mutations within the receptor sequence have been described that are able to increase the basal G protein activation level because the tri-dimensional structure of the receptor was probably modified by the amino acid exchange; these mutant receptors display constitutive activity (Lefkowitz et al., 1993; Pauwels and Wurch, 1998). In the present study, a mutation in a G protein exhibits a retrograde modulatory effect on the ligand binding properties of ₂ AR agonists. Recently, Grishina and Berlot (2000) showed that a chimeric G_s/i2 protein switching their ₃/₅ domains yielded an increased population of co-expressed ₂ AR in a high-affinity state (34%) compared with a wt G_s protein (19%) and a concomitant increase in the isoproterenol high- and low-affinity dissociation constants. These results also suggest, for another G protein domain, a modulatory effect on GPCR/G protein interactions. In contrast, the affinity constant of the antagonist yohimbine or the agonists clonidine and ()-epinephrine were either not decreased or maximally 2-fold decreased between the _2A-AR:G_i1Cys³⁵¹Gly and _2A-AR:G_i1Cys³⁵¹Ile fusion proteins (Jackson et al., 1999). The apparent absence of effect on agonist binding for the Cys to Gly and Ile mutations of the 4 C-terminal residue in the G_i1 portion of the fusion proteins may be due to the use of a labeled antagonist as a radioligand instead of an agonist. Other explanations may be: 1) the constrained interaction between the _2A-AR and the mutant G_i1Cys³⁵¹Gly/Ile proteins due to the fusion process, which may restrict the flexibility of the G_i1 protein partner, thereby masking effects that are uncovered by the co-expression experiments presented here, and 2) a weaker influence of the 4 C-terminal position of the G_i1 protein compared with the 3 Glu residue detailed here toward _2A-AR states.

In conclusion, the present data highlight a critical role for the C-terminal portion of the G_o protein in the modulation of _2A-AR states and the particular involvement of the third amino acid away from the G protein C-terminal extremity to determine the transition from a partial to efficacious agonist or antagonist at the _2A-AR. A retrograde modulatory effect of the G_o protein on the _2A-AR agonist binding site(s) is hypothesized, which probably involves transmission of the mutation-induced conformational change from the G_o protein to the ligand-bound _2A-AR.